Re: [PATCH 2/2] Flag and skip over packfiles known to be invalid.

From: Junio C Hamano <junkio@cox.net>
Date: Feb 2, 2007, 08:34 UTC
Message-ID: <7virelhsae.fsf@assigned-by-dhcp.cox.net>
In-Reply-To: <20070202080010.GB21094@spearce.org>

"Shawn O. Pearce" <spearce@spearce.org> writes:

> If we've already looked at a packfile and determined it isn't
> valid/usable as a pack, we shouldn't try to use it again in
> the future either.  This avoids multiple error messages from
> the same packfile.

I haven't looked your patch very closely, but two successive "git-repack -a -d" would produce the packfile under the same name, but with different parameters (--window or --depth), the old .idx and new .pack may not match.

Can this be possible (time flows from top to bottom):

	git-repack -a -d		somebody else

	produces a new .pack/.idx

        mv new .pack to objects/pack

					mmap .idx (old)
					finds .pack does not match

	mv new .idx to objects/pack
	
Is this something we care about?